RMG

RMG provides thermochemical, kinetic, solvation, and transport parameters for chemical reaction modeling using curated libraries, group additivity schemes, and machine learning predictors.

Key Features:

• Thermochemistry Estimation: Contains 45 thermochemical libraries (4,564 entries) and applies group additivity with nine correction types across 1,580 curated groups; includes a graph convolutional neural network trained via transfer learning from over 130,000 Density Functional Theory (DFT) calculations to 10,000 high-quality values.
• Kinetics Libraries and Rate Rules: Provides 92 kinetic libraries (21,000 reactions) and rate rule schemes for 87 reaction classes trained on 8,655 curated reactions.
• Solvation and Transport Models: Includes solvent–solute correction schemes with tabulated data for 195 solvents and 152 solutes, group additivity for arbitrary solutes, and estimators for transport properties.

Scientific Applications:

• Chemical Kinetic Modeling: Supplies parameters for constructing, validating, and analyzing reaction mechanisms and kinetic systems.

Methodology:

RMG integrates curated experimental and computational datasets with group additivity methods, rate rule training, and graph convolutional neural networks trained through transfer learning from large Density Functional Theory (DFT) datasets to predict thermochemical, kinetic, solvation, and transport properties.